Certainly; with a high enough dose, the subject would die.
That aside, the finding is interesting. Based on the summary, I thought that it might just be helping the subjects get closer to the ideal level of psychological arousal for what is probably a simple, routine, and possibly slightly boring task. However, the article states that the subjects were given the pills after having been shown the images, not before, in order to control for that possibility.
There is still one alternative explanation that I can see to a direct chemical effect of caffeine. For the subjects given caffeine after doing the first task in the experimental setting, an association was formed between the setting and getting caffeine. When the subjects returned to do a similar task in the experimental setting again, they received a slight boost in psychological arousal in anticipation of receiving the caffeine. (This kind of effect is commonly seen with many drugs, although I don't know if 200mg of caffeine would induce the effect with a single exposure.) The increased arousal during the follow-up task could explain the increased performance. If they wanted to control for it, one way would be to administer the follow-up task in a different environment than the one in which they did the first task, thereby reducing the impact of any associations with the original setting.